Technologies for Healthy Independent Living coincides with our goal to test and further develop inexpensive wireless sensors with communication and analysis platforms to monitor everyday activities, such as walking and exercise, in disabled persons. Our Medical Daily Activity Wireless Network (MDAWN) aims to enable researchers, trialists, and clinicians to acquire reliable data about daily physical activites in the home and community, rather than only during clinic and laboratory visits or by questionnaires. Novel machine-learning algorithms developed for each person from ankle accelerometers will identify the type, quantity and quality of exercise and walking on a continuous basis. In addition, we will test an inexpensive, instrumented, ergometric pedaler designed for disabled persons, called the UCFit, to promote home-based fitness exercise. The MDAWN and UFCFit sensors send data to an Android smartphone, then over the Internet to a UCLA server for analysis. We will nest further evaluation of the utility of these remote sensing devices within the real-world setting of a clinical trial of patients with recent disabling stroke,for whom formal rehabilitation often falls short of enabling functional walking and cardiovascular fitness. In a phase 2, randomized controlled trial, we will compare two levels of weekly telephonic feedback about daily performance, made possible for the first time by remote data acquisition. The high feedback group will receive everyday performance data that includes the number of bouts and minutes of pedaling plus rate/RPMs and forces used, as well as the number of bouts of walking, duration, and average speed and distance used in home and community. The minimal feedback group only hears about total daily UCFit exercise and walking time. After 4 months of encouraging up to four 30-min sessions per week of UCFit exercise, we will test for a 50% between-group difference in amount of daily exercise and walking time. Secondary outcomes examine change in level of fitness, walking speed, and physical functioning. The two groups then cross over to high or low feedback for 4 more months, before outcomes are reassessed. Then, no feedback is given and participants are tested for amount of exercise and fitness at 12 months. We will determine which form of feedback motivates better self-management to optimize fitness and walking. We will gather unique data about mobility in the transition from rehabilitation hospital to home; quantify gait training durin usual care; assess walking skills under more complex conditions than in a laboratory; and test the responsiveness of sensor data as a ratio scale outcome measurement during a time of expected gains. This study will be the first comprehensive demonstration of mHealth remote monitoring and outcome measures of daily physical activity!